Another concept that needs to be well understood before you start indexing -- perhaps the most important concept -- is how the index is going to be used. To get this clearly in mind, you can think of how you yourself use an index to find things in a large document. Traditionally, indexes are based around THINGS being discussed in the document that has been indexed -- NOUNS. If you had a cook book, you might want to be able to quickly find every recipe that uses chocolate (or any other ingredient that appears in the book). You might remember that a recipe you want to find again mentioned George Washington, so you would look under "Washington, George". (Every proper noun should be indexed, and if it is the name of a person, the index entry lists the last [family] name first.) You might want to find all recipes that are appropriate for a microwave oven, so you would search for "microwave" in the index. You might want to make a frosting for your cake, and remember having encountered a recipe called "Seven-Minute Frosting", so you might look under "Frosting" and find:
frostings
prune: 127
sea foam: 113
seven-minute: 53
snowy boiled: 276
These are specific frostings under the context entry "Frostings", or you might look under
Seven-Minute Frosting: 53
(Note that the other frostings above would each have their own individual level-1 index entries as well.)
If you remember the official recipe name, certainly THAT should appear in the index, possibly with different wordings in case you remember only one or two of the nouns or important words in the title.
The end result is a (hopefully thorough) table of terms and references (page numbers or links in computer documents) that predict in advance how most readers will think about (or remember) the NOUN you are referencing. They might think about it directly, or they might think about it in its context in the topic where it is discussed. And both the NOUN and its context may have alternate wordings! The number of combinations can and does add up rapidly.
And indeed, for a thorough index, a single paragraph in a detailed reference document commonly has in the range of 10-40 index entries.
